Effect of Mineral Particle Size on Froth Stability

Materials

A synthetic ore was prepared using pyrite and talc as the hydrophobic components and quartz as the hydrophilic component. Pyrite was obtained from Wards Scientific and quartz was obtained from Consol Glass (Pty). It was found necessary to use talc, a naturally floating gangue mineral in order to aid in stabilising the froth. The synthetic ore used contained 1.2% talc, 5% pyrite and 93.8% quartz.

A polypropylene glycol ether frother, Dowfroth 250, was used at a constant dosage of 100 ppm. A high frother dosage was necessary to run all the experiments in the bench scale continuous column under the same conditions, otherwise the froth volume obtained with coarse particles would be insufficient to get an acceptable mass flow to the concentrate. The glass wall does not offer a stabilising effect as it is hydrophilic and drainage was high at the surface. Potassium amyl xanthate (PAX) was used as collector at 25% surface coverage of pyrite, corresponding to a pyrite contact angle of 46°. The collector surface coverage was based on the BET surface area of the pyrite and assuming the cross-sectional area of the thiol head group of xanthate to be 28.8 Å (Grano et al. 1997). The collector dosage was chosen to give an intermediate particle hydrophobicity that would generate some pyrite recovery, but not high enough to destabilise the froth.

Sample preparation

Pyrite samples were pulverised using a Sieb mill (Ferguson Industrial Group). The pulverised material was then sieved into −25 μm, +25–75 μm, +75–150 μm and +150–300 μm size fractions. Samples were stored under nitrogen at −30°C to minimise mineral surface oxidisation. Talc samples were also prepared in the same size fractions.

Quartz samples were milled using 1 kg capacity Eriez Magnetics mill and sieved into the same size fractions as the pyrite and talc samples. All milling and screening were done on dry samples.

Mineral analysis

Quartz was assayed for iron impurities (that might interfere with the pyrite assays) using microwave digestion and atomic absorption spectroscopy (AAS), and the amount of Fe was found to be negligible. Concentrates and tails from flotation tests were assayed for Fe and Mg using microwave digestion and AAS.

Froth stability from non-overflowing stability column

The non-overflowing stability column is a PMMA column of internal diameter 9.3 cm and was used to study froth growth rates at a fixed air flow rate of 4 L/min, giving a superficial air velocity of 0.98 cm/s. 2 Litres of slurry was used for each experiment at 25% pulp density. The column has a P2 glass frit (of pore size 40–100 µm) for the generation of bubbles and an overhead stirrer to keep solids in suspension and facilitate bubble-particle collision. Froth stability tests were run on the same material that was subsequently floated using the continuous flotation column (Section 2.5).

Dynamic froth stability factor, ∑, is defined as (Bikerman 1973):. (1) where H_max = the maximum height attained by the froth; A = Column surface area and Q = Air flow rate. A dynamic froth stability factor was calculated for each experimental condition.

Continuous bench-scale column flotation tests

Flotation experiments were carried out using a 0.3 L (pulp section) bench-scale flotation column, operated continuously. The flotation column and its set-up is shown in Figure 1. The bench-scale flotation column was designed using PMMA as material of construction for the collection zone and glass for the froth zone, since bubbles tend to stick to the PMMA. The collection and froth zones are completely separable from each other. The collection zone has an internal diameter of 3.6 cm and length of 30 cm. It was necessary to have a froth zone of variable lengths i.e. 3, 7 and 11 cm so as to enable the calculation of the froth recovery using the variable froth depths technique, a method used by several authors (Laplante et al. 1983; Feteris et al. 1987; Vera et al. 1999). Since the froth height needed to be varied a detachable launder positioned on top of the froth zone glass column was used. The system was allowed to reach steady state in closed circuit for about 10 min, thereafter, three concentrates and three tails were collected in open circuit at 30 s intervals. 7 kg of slurry was prepared from the synthetic ore for each flotation test. The slurry feed flow rate used was 580 g/min at a residence time of 1 min and the superficial gas rate was 0.98 cm/s. The same pulp density and reagent conditions were used as for the froth stability column.

OPEN IN VIEWER

Figure 1

Setup and operation of the bench-scale continuous flotation column.

Bubble burst rate as a froth stability measure

The technique used in this paper involves taking videos during the flotation process from the top of the concentrate launder and using VirtualDub, a free online software, to analyse the bubble burst rate. By scrolling through the videos frame by frame, the total number of bubbles bursting per unit time for a given number of frames is counted and recorded and this represents the froth stability. 1/(bubble burst rate) defines the time it takes for a given number of bubbles to burst.

Representation of particle size data

Average particle size

The experiments were conducted using discrete size classes as well as mixtures of these size classes at two ratios, 1:4 (20% and 80%) and 4:1 (80% and 20%) as shown in Table 1. The weighted geometric mean was chosen to represent the average size of the particles in subsequent analysis.

OPEN IN VIEWER

Table 1

Particle size classes used in the experiments.

Percentage particles by mass from each size class (%)				Geometric mean size (µm)
−25 μm	25–75 μm	75–150 μm	150–300 μm
80	20			15.59
80		20		28.14
20	80			36.37
	100			43.30
80			20	49.35
	80	20		55.85
	80		20	77.07
20		80		86.58
	20	80		93.51
		100		106.07
		80	20	127.28
20			80	171.44
	20		80	178.37
		20	80	190.92
			100	212.13

Feed versus concentrate particle size

The graphs that follow use the feed particle size rather than concentrate particle size to analyse the data. Naturally, the particles that will affect the froth stability are those that report to the froth phase. These may not necessarily be of a similar size to the feed particle size. In the first phase of data analysis, froth stability was analysed as a function of feed particle size, rather than concentrate particle size. On a full-scale plant operation or in ore characterisation, feed particle size is generally a more easily measured property than concentrate particle size, and it is a more useful property for ore characterisation. Figure 2 illustrates the relationship between feed particle diameter and concentrate particle diameter. It is clear that the particles reporting to the concentrate have a slightly smaller average diameter, as would be expected. However, they follow the same linear trend and plots of froth stability versus either feed or concentrate particle diameter follow the same trend and are almost identical. It was decided to analyse the data in terms of the feed particle diameter since this is a parameter that is more useful as a predictive tool.

OPEN IN VIEWER

Figure 2

Feed particle diameter versus concentrate particle diameter. The line represents the xy function.

Dynamic froth stability and particle size

Figure 3 shows the relationship between particle size and dynamic froth stability using a non-overflowing stability column. Results show a power law dependence of the froth stability on particle size. The froth was most stable at small particle sizes and least stable at large particle sizes. Dynamic froth stability varied between 68 s, for the finest particle sizes and 7.7 s for the coarsest particles. There was a steep initial decrease in the froth stability as the particle size increased up until about 50 µm, where after the dependence of froth stability on particle size became less pronounced.

OPEN IN VIEWER

Figure 3

Effect of feed particle diameter on dynamic froth stability.

It was observed visually in the stability column that fine particles formed fine bubbles that grew steadily up the pulp-froth interface. The froth was very stable and carried a high amount of water and solid particles. In contrast, coarse particles formed froths with larger bubbles that were very unstable. The froth had a low concentration of particles. As such, the froth height obtained for coarse particles was low.

It is significant that both discrete size fractions as well as 1:4 and 4:1 mixtures of particle sizes all fall on the same curve. The particles stabilise the froth to the same extent whether they are, for example, a mixture of small and large particles or a narrow size range of medium sized particles. Dippenaar (1982b) found that the rate determining step in the rupture of bubbles by hydrophobic particles of different sizes was the rate at which the thinning of the films occurred. Large particles ruptured films faster since less film thinning was required before the particle could bridge the film and cause migration of the three-phase boundary lines to the same point. It may have been expected that a single large particle would be sufficient to bridge the film and cause rupture. However, it is clear from Figure 3 that all particle size combinations fall on the same relationship.

It is significant that froth stability rises sharply for a feed size less than 50 μm. This raises the question of whether the froth is stabilised by entrained particles, since the amount of these would be expected to rise sharply at a particle size of less than 50 μm, or by naturally floatable particles that are attached at the interface, or by a combination of the two. Entrainment is the non-selective recovery of particles in the water which reports to the concentrate. The proposed mechanisms of film stabilisation incorporate particles attached at the air–water interface as well as particles with complete wettability that reside within the films, as discussed in the introduction.

Figure 4 shows the amount of total solids reporting to the concentrate as a function of the feed particle size. In addition, the total solids have been separated into hydrophilic material (the quartz), which is presumed to be recovered mostly by entrainment and the hydrophobic material (the pyrite and talc), which is recovered mostly by true flotation and is attached at the air–water interface. It is evident from Figure 4 that the amount of hydrophilic quartz reporting to the concentrate increases sharply below a feed size of 50 μm. Although there is some scatter in the data, little quartz is recovered to the concentrate above a feed size of 50 μm. This is in line with many previous studies that have shown entrainment to be significant below a particle size of 50 µm (Smith and Warren 1989). The recovery of hydrophobic material follows a gradual, linear increase as particle size decreases. Thus, it is evident that the dramatic increase in froth stability below a particle size of 50 μm can be attributed to the large increase in the amount of hydrophilic particles reporting to the froth. This suggests that the drainage of the thin films is affected by the presence of hydrophilic particles within the films and that this has a large effect on the froth stability. The drainage is affected by an increase in viscosity of the liquid making up the films, which will increase with decreasing particle size and increasing particle mass or area. In addition, foam stability may be described by the pressure gradient between bubbles and interfilm fluid, or the capillary pressure. The thin liquid films between bubbles are stabilised by the maximum capillary pressure, . This is the pressing force required to bring two bubbles to coalescence. This observation was made by Denkov et al. (1992) who considered theoretical liquid films formed by a particle monolayer pressed between two emulsion interfaces. Utilising the works of Denkov et al. (1992)., Kaptay (2006) derived an expression for as:. (2) where = particle radius, ℴ = interfacial tension, = contact angle, and = theoretical packing parameter, which is a function of particle concentration and packing on the capillary pressure.

OPEN IN VIEWER

Figure 4

Flowrate of total solids (blue diamonds), hydrophilic material in the form of quartz (orange squares) and hydrophobic material in the form of pyrite and talc (grey triangles) as a function of feed particle diameter.

Equation 2 dictates that a froth system will only be stable if the maximum capillary pressure is positive and is much larger than the combined sum of all forces (centrifugal, electric or magnetic fields, etc.) That seek to collapse the thin liquid films between the bubbles. The maximum capillary pressure is seen to be inversely proportional to the particle size, which shows that smaller particles are more effective at preventing drainage and stabilising foams. Thus, it is expected that the main mechanism stabilising froths with feed particle sizes of less than 50 μm is the increased capillary pressure gradients and viscosities of the thin films. Above 50 μm, there is a mixture of both hydrophilic and hydrophobic particles reporting to the froth phase. However, the combined mass flow of these components is much reduced above about 50 μm in size.

As expected, as the amount of solids in the concentrate increases so the froth stability also increases as depicted in Figure 5. This shows that an increase in particle recovery to the concentrate launder leads to an increase in the stability of the froth. In essence, froth stability is dependent on the amount of solid particles present in the froth.

OPEN IN VIEWER

Figure 5

Relationship between dynamic froth stability and concentrate solids flow rate.

Dynamic froth stability and particle surface area

It might be postulated that the particle size is significant in stabilising a froth in terms of the surface area that it represents. The froth stabilisation mechanism of both hydrophobic and hydrophilic particles can be related to their area. An inverse relation exists between particle size and surface area, i.e. (3)

Where d = particle diameter; α = shape factor; A_BET = specific surface area as determined by the BET technique; and = particle density.

The reciprocal relationship between particle size and specific surface area (depicted in Equation 3) is used in Figure 6 to show the relationship between the dynamic froth stability and particle specific surface area. This shows that the froth stability is linearly dependent on the specific surface area of the particles.

OPEN IN VIEWER

Figure 6

Relationship between dynamic froth stability and reciprocal particle size, representing surface area.

The linear relationship shown in Figure 6 represents a useful tool with which to predict froth stability on the basis of feed particle size. The relationship seems to hold irrespective of whether the froth is stabilised by hydrophilic or hydrophobic particles. It is expected that the gradients of these lines will be dependent on the hydrophobicity of the mineral particles. If these linear relationships can be calibrated for a particular ore type, this may be a useful tool for predicting froth stability as a function of grind for full-scale systems.

Comparison of froth stability versus particle size using different systems

The non-overflowing froth stability column provides a practical and simple method for determining froth stability. In addition, it is a potentially scalable measure, which is the subject of ongoing research in our laboratories at the Centre for Minerals Research. However, it does not provide any metallurgical data, which is why the bench-scale continuous column flotation cell was designed to simultaneously collect froth stability and metallurgical data. Two froth stability measures were used in the continuous column flotation cell: water recovery and top-of-froth bubble burst rate. The relationship between froth stability, top-of-froth bubble burst rate and water recovery is well known in flotation literature (Hadler and Cilliers 2009; Morar et al. 2012; Neethling and Brito-Parada 2018). The greater the film stability, the lower the burst rate and the more stable the froth is. It naturally follows that as more bubbles burst, they release the water trapped in the Plateau borders and less water is recovered to the concentrate. It is also well known that a more stable froth will hold more water in the thin films and Plateau borders (Triffett and Cilliers 2006). In the case of froth stability prediction by the bubble burst rate technique, the reciprocal of the bubble burst rate (i.e. 1/bubble burst rate) defines the time it would take for a given number of bubbles to burst. The greater the bubble bursting phenomena at the top of the froth, the less time it would take for a given number of bubbles to burst. When bubbles burst, particles and liquids in the Plateau borders are released and drain freely under gravity, although there is a possibility that they could still be captured by a rising froth.

Figure 7 shows the froth stability measures of water recovery and bubble burst rate in the continuous column, with the non-overflowing column dynamic froth stability shown for comparison as a function of inverse particle size. It is apparent that, with some scatter, all measures conform to a linear relationship. Thus, the same mechanisms that govern the relationship between particle diameter and froth stability in the stability column also govern the relationships in the continuous column. It should be noted that in both pieces of equipment, the same superficial air velocity of 1 cm/s was maintained. The air flow rate will govern how much water reports to the froth and subsequently to the concentrate launder. Likewise, this should also govern the growth rate of the froth since the frother dosage was held constant in both devices. The major difference between these two pieces of equipment is the fact that the flotation column cell is an overflowing, continuously operated cell, whereas the froth stability column is a non-overflowing, batch cell. It was not known whether the froth stability column would conform to the continuous, overflowing process because there may come a point where the gravitational forces would overcome the upward flow of air, water and particles and froth growth will be curtailed. This would give a lower than expected froth stability measurement. However, as shown in Figure 7, there is correlation between the three measurement types.

OPEN IN VIEWER

Figure 7

Froth stability measures in different systems as a function of inverse particle diameter. Froth stability measures were: (a) Water flow rate in the continuous column (blue diamonds); (b) Bubble burst rate in the continuous column (grey triangles); and (c) Dynamic froth stability in the non-overflowing column (orange squares).

It should be noted that both devices are dynamic processes, meaning that liquid is reporting to the froth phase continuously while drainage is occurring simultaneously. This is as opposed to a static process, where there is no air input and only drainage is occurring. The froth growth height, bubble burst rate and water recovery will all be driven by these two processes: water and particles reporting into the froth, while simultaneously draining. Liquid content of the froth at the surface affects the liquid films at the top of the froth (Neethling and Cilliers 2003). Wet froths have smaller bubbles and therefore contain more water, whereas dry froths have larger bubbles and contain less water (Neethling and Cilliers 2003; Morar et al. 2012). Since bubble coalescence causes materials (particles and liquid trapped within the lamella) to detach from the bubbles, the rate of coalescence will be higher in a dry froth than in a wet froth. The bubble burst rate at the top of the froth is related to the incoming bubble size as well as pulp conditions. It is expected that the wet froth will experience less surface bubble bursting phenomena than the dry froth. Less surface bubble bursting phenomena (since it takes a longer time for a given number of bubbles to burst) implies more water flow over the weir and therefore a higher water recovery.

Particle size, froth recovery and flotation performance

Flotation performance was monitored concurrently with the froth stability measurements. Figure 8 shows that the pyrite recovery follows the similar type of power law dependence on particle size as the froth stability data. This recovery is a combination of the pulp phase and froth phase recovery. According to the literature, pulp phase recovery follows the classic dependence on particle size first put forward by Gaudin et al. (1931), where smaller particles and larger particles are more difficult to recover than intermediate sized particles. These recovery losses occur mainly due to limitations associated with bubble-particle adhesion and collision rates, detachment and buoyancy (Kohmuench and Mankosa 2012). Interestingly, Figure 8 shows that there is not a decrease in the pyrite recovery at smaller particle sizes. This may indicate that the stability of the froth zone is more important to the recovery of pyrite than the pulp zone. The expected decrease in pyrite recovery in the pulp zone at smaller particle sizes is offset by the higher froth stability at those particle sizes, which aids in the recovery of pyrite.

OPEN IN VIEWER

Figure 8

Effect of particle size on pyrite recovery.

Figure 9 shows the valuable mineral rate constant as a function of froth stability. This serves to show the strong role that froth stability plays in valuable mineral recovery. With an increase in froth stability, there is a concomitant increase in valuable mineral recovery, as reflected by an increase in the pyrite rate constant. This shows the value of being able to predict froth stability for different particle properties.

OPEN IN VIEWER

Figure 9

Variation of froth stability with pyrite rate constant.